Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Heckman and Oaxaca

From   [email protected]
To   [email protected]
Subject   Re: st: Heckman and Oaxaca
Date   Thu, 28 Oct 2004 07:12:03 -0700


There is no problem with dummies in a Heckman equation and "decomp."
decomp uses the mean of the variable, in this case the proportion of the
sample with dummy variable equal to one, and multiplies it by the
respective coefficient.  To see the details of the calculations, use can
use "himod, ds" and "lomod, ds."  this will give you the contribution of
each variable to the total endowment and the total residual.

It is possible to get high percentages (-301.0).  It means that the group
that you are using as the "low" or discriminated group actually has
advantage over the "high" group.  The commend, "himod, ds" or "lomod, ds"
will reveal what variables are the cause of high percentages.


                      "Natalia Ferreira"                                                                                               
                      <[email protected]>             To:      [email protected]                                        
                      Sent by:                          cc:                                                                            
                      owner-statalist@hsphsun2.         Subject: st: Heckman and Oaxaca                                                
                      10/27/2004 05:04 p.m.                                                                                            
                      Please respond to                                                                                                

Hi everybody

    I�m estimating a Heckman wage equation to migrants and not migrants,
after all I want to compare them by a Oaxaca decomposition, but I have a
question: there is any problem if my Heckman equation have some dummies?
"decomp" command deal with the dummies variables? Does it take means of
and ones? The results of my "decomp" give me discrimation % of -301.0
The % should not be between 0-100? The results are below:

Summary of decomposition results (as %)
Amount attributable:             |    -10.3
- due to endowments (E):         |      5.2
- due to coefficients (C):       |    -15.5
Shift coefficient (U):           |     11.6
Raw differential (R) {E+C+U}:    |      1.3
Adjusted differential (D) {C+U}: |     -3.9
Endowments as % total (E/R):     |    401.0
Discrimination as % total (D/R): |   -301.0

Thanks for the help, Natalia B.
P�s Gradua��o IPE/USP

MSN Messenger: instale gr�tis e converse com seus amigos.

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index